For quite a long time I have been working on the relationship between colour and meaning. A lot of very good research has been carried out over the last few decades in this area. Typically, in this research, people are shown a colour and asked to respond what they think about it on a bi-polar scale; for example, is it warm or is it cool. Or, is it modern or is it old-fashioned. This research is very nice but in my opinion we should turn the problem on its head. Rather than asking what we think about a colour we should ask which colours do we associate with warm, which colours do we associate with modern, which colours do we associate with happy. There are several reasons for this. One is because I would like the research to be useful to, for example, designers. Designers don’t typically start with a colour and wonder what that colour represents. Rather, they might start with a brief that includes some concepts, such as modern and financial and think start to explore which colours might represent (or communicate) those words or concepts. It’s great to do this sort of research in the lab. However, I would really like to be able to generate a colour palette for any word or concept and it is clear that we can’t run costly and laborious lab experiments for every word (and for every word in every language actually).

I worked with a major paint company to automatically generate colour palettes for words using large-scale internet scraping following by machine learning. We have a pilot website with results for a few words. It looks like this.

You can see this pilot website here.

Another way to get lots of data about this topic is using crowd sourcing. We have also been doing this and you can read our latest paper about this in the Journal of the International Colour Association. Details of this paper are shown below: